The Political Nature of Entrepreneurship in Developing Countries

Experimental Evidence from Tunisia and Senegal

Robert Kubinec

University of South Carolina

Abhit Bhandari

Vanderbilt University

Sekou Jabateh

University of California Berkeley

Hamza Mighri

International Monetary Fund

June 28, 2024

Introduction

What Are Political Connections For?

  • We know that sometimes political connections hurt economic growth by encouraging corruption (pretty much everywhere except Japan).

  • If connections can be so important for firms in developing countries, how do young entrepreneurs gain access to such connections if they do not have them?

  • To what extent do political connections as a fixed resource endure over time?

No Meme Slide This Time

Case Selection

  • We look at young people in Senegal and Tunisia who belong to the population of would-be entrepreneurs.

  • We focus on young people because:

    • Their connections are largely fixed by family relationships.

    • Their careers are malleable.

    • We can track them over time to see how they make decisions about careers.

Case Selection

  • Senegal and Tunisia are rated as top 10 countries in the world for interest in entrepreneurship (Global Economic Monitor).

  • Both measure poorly on indices of corruption in government (Senegal #72/180 and Tunisia #85/180).

  • Both are former French colonies with heavy-handed bureaucratic regulation that can make connections very important to ease red tape.

Hypotheses

Hypothesis 1: An increase in a young entrepreneur’s political connections should cause an improved self-assessment of the possibility of engaging in entrepreneurship.

Hypothesis 2: Increased interactions with government officials will result in higher self-reported political connections among potential entrepreneurs.

Multi-Stage Research Design

  • Stage 1: Recruit quasi-representative sample of young people (ages 18 - 30) via Facebook ads. Screen for interest in entrepreneurship or government careers.

  • Then invite this sample to a second survey with a larger mobile credit (~$10 USD).

  • Second survey contains a conjoint experiment (4 tasks) compares two hypothetical entrepreneurs and includes connections (father’s profession, membership in ruling party) as attributes.

  • In total, we collected 609 responses in Senegal and 501 in Tunisia (N for conjoint is 8,156).

Conjoint Example

Multi-Stage Research Design

  • Stage 2: Run in-person field experiment by inviting random sample of survey respondents to entrepreneurship training.

    • Respondents given incentive to cover travel costs.

    • We hired entrepreneurs to design and run workshops in both Tunisia and Senegal.

      • Government officials from business development agencies invited to present.
  • In total 83 participants; all others are controls.

Multi-Stage Research Design

Outcomes:

  1. How likely are you, alone or with others, to start a new business, including any type of self-employment, within the next three years? (Ordinal response)
  2. Over the past year, did you, alone or with others, try to start a new business? (Yes/No)

Multi-Stage Research Design

  • Finally, we re-contacted respondents 1 year (Tunisia) and 6 months (Senegal) after completion of the training. All respondents filled out the complete initial survey instrument.

  • Recontact rates were 57% for Tunisia and 78% for Senegal (attrition generally occurred in the control group).

Inference Problem

Descriptives

Examining Political Connection Relationships

Conjoint Results

Field Experiment Results

Total Effect of Treatment on Entrepreneurship

Indirect Effect of the Treatment Via Connections

Mechanism: Interactions with Officials

Long-term Effects on Owning Businesses

Did We Convince Them Connections Don’t Matter?

Conclusion

  • It does seem that we can manipulate perceived connections through meetings with government officials.

  • Increasing connections does seem to contribute to improved odds of perceived interest in entrepreneurship–and probably success as well.

  • Effects of the treatment on interest/attempts last over the long-term, but not the political connections component.

Qualitative Analysis of Connections

  • We also asked people to describe the connections using open-ended text responses:

    • Our relationship isn’t close at all because he thinks I need something from him when I’m not in I only think about my future so I don’t count on him
    • We see each other no more than twice a year.
    • He’s my namesake he was a former Minister of State
    • None strangely, he doesn’t even know me but I know he is a distant family relation
    • Gives a lot of importance to me, a welcoming person, keep your head up
    • Professional relationship in the context of finding a job
    • He’s my father’s uncle. He was political adviser to the President of the Republic
    • He’s like a father to me because he’s a friend of my uncle
    • I am a member in a political party currently (pdl)

Connections and Income

Connections and Public Sector Interest

Change in Political Connections Over Time

Indirect Effect of the Treatment Via Connections (Details)

Effect Outcome Mediator Type 5% Median 95%
Direct Intentions General 0.029 0.063 0.099
Direct Start Business General 0.111 0.175 0.248
Direct Intentions Parliamentary 0.051 0.089 0.124
Direct Start Business Parliamentary 0.120 0.178 0.261
Indirect Intentions General 0.001 0.005 0.009
Indirect Start Business General 0.000 0.002 0.006
Indirect Intentions Parliamentary -0.003 0.000 0.003
Indirect Start Business Parliamentary -0.001 0.000 0.001
Proportion Mediated Intentions General 0.066 0.181 0.347
Proportion Mediated Start Business General -0.005 0.037 0.087
Proportion Mediated Intentions Parliamentary 0.029 0.103 0.203
Proportion Mediated Start Business Parliamentary -0.019 0.021 0.057
instrument Intentions General 0.215 0.263 0.311
instrument Start Business General 0.269 0.330 0.390
instrument Intentions Parliamentary 0.225 0.272 0.323
instrument Start Business Parliamentary 0.277 0.334 0.389
med_on_out Intentions General 0.005 0.014 0.022
med_on_out Start Business General -0.001 0.006 0.015
med_on_out Intentions Parliamentary 0.003 0.010 0.019
med_on_out Start Business Parliamentary -0.003 0.004 0.010
treat_on_med Intentions General 0.164 0.355 0.545
treat_on_med Start Business General 0.157 0.375 0.610
treat_on_med Intentions Parliamentary -0.231 0.010 0.241
treat_on_med Start Business Parliamentary -0.229 0.033 0.308

Long-term Effects (Details)

Outcome Type 5% Median 95%
Employ People Direct Effect (path $d$) -24.200 -7.115 8.191
Employ People Mediation General Connections (path $m$) -2.231 0.279 3.158
Employ People Mediation Parliamentary Connections (path $m$) -5.413 -2.514 0.536
Own Business Direct Effect (path $d$) -0.012 0.106 0.246
Own Business Mediation General Connections (path $m$) 0.007 0.025 0.049
Own Business Mediation Parliamentary Connections (path $m$) 0.002 0.019 0.037
Pay Salary Direct Effect (path $d$) -0.370 -0.197 0.032
Pay Salary Mediation General Connections (path $m$) -0.003 0.027 0.052
Pay Salary Mediation Parliamentary Connections (path $m$) -0.006 0.029 0.063
Quit Business Direct Effect (path $d$) 0.058 0.212 0.373
Quit Business Mediation General Connections (path $m$) 0.011 0.031 0.048
Quit Business Mediation Parliamentary Connections (path $m$) 0.018 0.041 0.064

References